Ux Writing

Deep UX writing workflow—voice, clarity, error and empty states, forms, accessibility of text, localization hooks, and collaboration with design. Use when po...

Security Scan

An instruction-only UX writing workflow that is internally consistent with its description and does not request unusual permissions, installs, or credentials.

Purpose & Capability

The skill's name and description match the SKILL.md content: a multi-stage UX writing workflow. It does not request binaries, credentials, or config paths that would be unrelated to UX writing.

Instruction Scope

Runtime instructions ask for screenshots, current copy, and metrics (support tickets, drop-off). Those inputs are reasonable and relevant to the stated purpose, but they can contain sensitive or PII-containing material — the skill does not provide handling/retention rules, so callers should redact or authorize access appropriately.

Install Mechanism

No install spec and no code files (instruction-only). This minimizes disk writes and arbitrary code execution risk.

Credentials

The skill declares no environment variables, credentials, or config paths. Nothing requested is disproportionate to a UX writing workflow.

Persistence & Privilege

always is false and the skill does not request persistent platform privileges or modify other skills. Autonomous invocation is allowed by default but is not itself a red flag here.

Assessment

This skill appears coherent and low-risk, but be mindful of the input data you provide: - Redact or anonymize screenshots, support tickets, and any logs that may contain PII or secrets before uploading. - Do not paste full production logs, stack traces, or credentials; share only the copy/context needed for rewriting. - If error messages, login flows, or account state are involved, coordinate with security/legal before sharing examples that could reveal account enumeration behavior. - For localization work, validate that your i18n pipeline and pseudolocale testing are in place rather than embedding locale-specific assumptions in example strings. - Confirm retention and access policies for any artifacts the skill will consume (screenshots, metrics) so sensitive data isn't stored longer than necessary.

UX Writing (Deep Workflow)

UX writing is interface design with words: reduce cognitive load, prevent errors, and build trust. It is not marketing polish bolted on at the end.

When to Offer This Workflow

Trigger conditions:

• Confusing errors, high drop-off flows, empty states that feel broken
• Inconsistent terminology across product
• Accessibility review flags unclear labels or verbose instructions
• Preparing voice & tone guidelines for a design system

Initial offer:

Use six stages: (1) context & users, (2) voice & tone, (3) clarity & structure, (4) errors & recovery, (5) forms & validation, (6) a11y & localization. Ask for screenshots, current copy, and metrics (support tickets, drop-off).

---

Stage 1: Context & Users

Goal: Copy matches mental model and emotional state.

Questions

1. User goal on this screen—primary action?
2. Stress level: billing, security, health—tone adjusts
3. Expertise: first-time vs power user—progressive disclosure

Constraints

• Character limits in UI components; legal must-review text

Exit condition: Scenario brief per screen or flow—not generic “friendly.”

---

Stage 2: Voice & Tone

Goal: Consistent personality with flexible tone by context.

Voice (stable)

• Principles: e.g., clear, respectful, confident, human—pick 3–4 and define anti-patterns

Tone (situational)

• Success: brief affirmation
• Error: calm, no blame; next step forward
• Empty state: invite action without condescension

Terminology

• Glossary: “Workspace” vs “Project”—one term per concept; align with engineering names users see in API/docs

Exit condition: Before/after examples for three contexts (success, error, empty).

---

Stage 3: Clarity & Structure

Goal: Scannable text—front-load meaning.

Practices

• Titles specific: “Payment failed” not “Something went wrong” (unless generic is truly unknown)
• Buttons use verbs: “Save address” not “OK”
• Sentence case per style guide; avoid ALL CAPS except acronyms
• Numbers/dates: user locale; relative time when helpful (“Updated 2 min ago”)

Microcopy hierarchy

• Primary message → secondary detail → tertiary learn more

Exit condition: Redundant words cut; one idea per sentence in critical paths.

---

Stage 4: Errors & Recovery

Goal: Users understand what happened and what to do next.

Structure

• What happened (plain language, no codes alone)
• Why (if known and helpful—not stack traces)
• What to do (steps, link to support, retry)
• Support path when stuck

Security

• Don’t leak whether an email exists on login if policy requires ambiguity—coordinate with security

Exit condition: Top 10 error states have rewritten copy + engineering alignment on truth of messages.

---

Stage 5: Forms & Validation

Goal: Inline help and validation text accessible and timely.

Practices

• Label every input; don’t rely on placeholder alone as label
• Errors associated programmatically (aria-describedby); announce on submit failure
• Password rules visible before typing when complex
• Success confirmation for destructive actions

Tone on errors

• Avoid shame (“Invalid input”) → neutral (“Enter a date after Jan 1, 2024”)

Exit condition: Form review checklist applied to highest-traffic form.

---

Stage 6: Accessibility & Localization

Goal: Text works for screen readers and translation.

Accessibility

• Alt text for meaningful images; decorative marked so
• Instructions not color-only; error identification not by color alone

Localization (i18n)

• No concatenated strings with word order assumptions across languages
• Punctuation and formality per locale—pseudolocale QA for overflow

Exit condition: String extraction friendly; no embedded HTML in strings without plan.

---

Final Review Checklist

• Voice/tone documented with examples
• Critical paths scannable; verbs on CTAs
• Errors: cause + next step + support path
• Forms: labels, validation, a11y association
• i18n-safe string patterns

Tips for Effective Guidance

• Read aloud—if awkward, revise.
• Pair with design: copy length affects layout; don’t fight the grid blindly.
• For AI products, clarify machine vs human responsibility in copy.

Handling Deviations

• Dense enterprise UIs: prioritize task efficiency over personality.
• Regulated industries: legal review loop—suggest plain alternatives, not legal advice.